行级别的值总和

Question

我有一个 SQL 数据作为

BugID  Request Date  Closed Date
--------------------------------
1234   11/1/2013     11/5/2013
2345   11/1/2013     11/2/2013
2435   11/2/2013     11/2/2013
6345   11/3/2013     11/5/2013
9364   11/5/2013     11/6/2013

我希望结果是

Date       Requested Bugs  Closed Bugs  Total Open  Total Closed
----------------------------------------------------------------
11/1/2013  2               0            2           0
11/2/2013  1               2            3           2
11/3/2013  1               0            4           2
11/5/2013  1               1            5           3
11/6/2013  0               1            5           4

是否可以通过 Transact SQL 实现？

Answer 1

这样的事情应该可以解决问题：

with allDates (date) as
(
    select requestDate from dbo.bugs
    union select closedDate from dbo.bugs
)
select date,
    (select count(*) from dbo.bugs where requestDate = a.date) as bugsRequested,
    (select count(*) from dbo.bugs where closedDate = a.date) as bugsClosed,
    (select count(*) from dbo.bugs where a.date between requestDate and closedDate) AS bugsOpen,
    (select count(*) from dbo.bugs where closedDate<=a.date) AS totalClosed
from allDates a

现场示例：http://sqlfiddle.com/#!6/de05f/15

Answer 2

您可以在 SUM/COUNT 函数中使用 CASE 语句，类似这样

SELECT   CAST([Request Date] AS DATE) AS [Request Date]
       , COUNT(*) AS [Requested Bugs]
       , COUNT(CASE WHEN [Closed Date] IS NOT NULL THEN 1 ELSE NULL END) AS [Closed Bugs]
       , COUNT(CASE WHEN [Closed Date] IS NULL THEN 1 ELSE NULL END) AS [Total Open]
FROM Table_Name 
GROUP BY CAST([Request Date] AS DATE)

我在这里考虑了一个封闭的列，它将指示哪些作业仍然打开，它可以是位字段或其他任何内容，但主要思想是向您展示如何在聚合函数中使用 CASE 语句只计算某些行。